The aging brain maintains the ability to learn while being affected by neuronal death, vascular abnormalities and connectivity loss. A possible explanation for this preserved function is the maintenance of neuroregenerative potential even in late-life. Strong emerging human in-vivo evidence indicates that the hippocampal formation becomes larger in response to physical activity (PA) interventions in well-functioning elderly in their late sixties. However, it is not known whether very old and frail individuals also respond to PA with hippocampal expansion and -if they do- whether changes remain beyond one year. Moreover, it is unknown whether hippocampal changes are specific for areas with neuroregenerative potential (the dentate gyrus), and the effects of PA on related cerebrovascular and functional brain measures. Addressing these gaps in knowledge in very old adults will expand our understanding of brain plasticity in very old age, specifically in response to intervention. We propose to address this gap in knowledge by examining whether the PA-related effects previously reported in the hippocampal formation of younger elderly exist in very old and frail adults for periods of time beyond 1 year, whether these effects are significant for the dentate gyrus, microcirculature and functional connectivity. To date, there have been two major barriers to answer these questions: 1) conventional non- invasive imaging cannot precisely measure the dentate gyrus; 2) very old adults are likely to be physically frail and unlikely to participate and adhere to a PA program for long period of times. This proposal addresses these barriers by applying non-invasive Ultra High-Field neuroimaging at 7Tesla in frail older adults who are already participating in an existing long-term randomized clinical trialof walking compared to health education (Successful Aging: SA). We have already acquired Ultra High-Field brain images at the time of randomization in 2010 in 65 ptc and have funds to acquire all brain MRI at the 2nd yr follow-up, for an estimated sample size of n=50 with MRIs pre and post-intervention and an additional n=30 with MRI at the post-intervention only. This is an application to process and analyze these data to: 1) Quantify the relationship of PA with Ultra High-Field brain markers (AIM1); 2) identify the localization of the PA-related changes in Ultra High-Field brain markers (AIM2); 3) Explore the role of Ultra High Field markers in explaining brain response to PA (AIM3).This proposal has a high chance of success because it applies previously validated neuroimaging protocols in response to a type of PA intervention that has also been previously validated in this special population. This is a unique opportunity to conduct this transformative study in a highly cost-effective design, leveraging existing data on Ultra High field brain MRIs, existing operational infrastructure for recruitment, retention, intervention, and clinical assessment of the ongoing LIFE study and the expertise of the LIFE leadership team, Co-I's of this proposal. Because of the high costs of RCTs, it is highly unlikely that this combination of highly advanced neuroimaging and type of intervention study will occur again.